Every homelab outgrows the browser bookmark folder, and Compass answers by building the front page for you, discovering your services automatically instead of asking you to maintain cards by hand. Point it at Docker, Kubernetes, Tailscale, Headscale, NetBird, DNS records, a JSON API, or a plain static list, and it renders everything it finds as a searchable directory with icons, descriptions, and tags filled in from a catalog when a source leaves them blank. Each card shows whether the service is actually up, using the Prometheus or Gatus monitor you already run, and pinned and recent services keep daily tools one click away. Opening a service reveals its connection details, source metadata, extra links such as runbooks and health endpoints, and panels you embed from Grafana. Markdown pages live beside the directory, so architecture notes and on-call runbooks sit in the same navigation, with live service cards and charts embedded directly inside the prose. A debug view shows per-source health, what each source returned, and why a service you expected is missing, while group-based access decides which sources each person can see. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache License 2.0 licensed.

For teams drowning in pipeline output nobody reads, Depictio turns workflow results into dashboards you can hand to a collaborator. A MultiQC tab answers the first questions about a sequencing run: how many reads survived trimming, how evenly they spread across samples, and whether base quality holds to the end of the read, all in one glanceable table. Filter by sample, habitat, site, or sampling date on the left and every panel on the page follows, from community composition cards counting distinct taxa and phyla to differential abundance views that show what was tested, what survives correction, and how large the effects are. Alpha diversity tabs summarise richness and evenness per sample with rarefaction curves, and a sampling sites map lets you click or lasso points to filter the whole dashboard by geography. Forty two ready-made templates cover nf-core pipelines such as ampliseq, atacseq, and airrflow, while outputs from Nextflow, Snakemake, Galaxy, or plain shell scripts feed the same builder. Even a classic tabular dataset becomes an overview of distributions, box plots, and threshold counts, so the same tool serves a wet lab and a data science team. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

One login for every internal tool, without handing your users another password to forget: Issuerd is a Keycloak-compatible identity provider that runs single sign-on for your whole stack from a single binary. Administrators manage realms, users, groups, clients, roles, client scopes, and protocol mappers in an embedded admin console, while end users get their own account console for passwords, passkeys, and two-factor settings. Authentication flows cover MFA, social login, and themable, localized login pages with per-realm language and email templates, so each audience sees a sign-in that looks like your product. Federation connects Samba AD, OpenLDAP, and Active Directory with user and group sync, password validation and write-through, plus Kerberos and SPNEGO for desktop single sign-on. For modern clients it implements DPoP token binding so stolen access tokens are useless without the holder's key, CIBA for approvals that happen on a second device, and token exchange patterns aimed at AI agents calling downstream services. Declarative YAML provisioning, a full management API, OpenAPI export, Prometheus metrics, and health probes round out day-two operations, and nodes scale horizontally behind any load balancer with no sticky sessions. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache-2.0 licensed.

Tasterr is a self-hosted discovery front end that learns what each person in your household actually likes and fills the home screen with movie and show rails picked for them. It reads the TMDB catalog in your chosen metadata language and merges it with your Plex, Jellyfin, Seerr, and Trakt accounts, so library titles, requests, and watch history shape every recommendation. Continue Watching pulls resumable movies, episodes, and next-up picks from Plex and Jellyfin into one rail, and Because You Watched rails spin off new suggestions from finished titles. Watched titles disappear from discovery by default, a Not interested button banishes titles for good, and My List keeps the maybes one tap away. When the household gathers, Something for Everyone Tonight blends member profiles into one shared shortlist, and Picks You Wouldn't Usually Watch deliberately wanders outside each profile. Requests flow into Seerr under the signed-in member's name, availability badges show where a title already lives, and admins tune regions, streaming services, rails, language, theme, and accent color. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.

Flopwire lets your AI coding agents message each other and search each other's history, so a Claude Code session on your laptop can ask the Codex session on your server what it already tried before duplicating the work. Agents see who is online and what they started on with peers, then reach each other mid-turn with send; a message arrives inside the recipient's running turn, or with its human's next prompt, and only your own live session can be woken deliberately. A device agent indexes transcripts from Claude Code, Codex, Devin CLI, and opencode within about a second; grep runs regex over every session, and search returns ranked matches naming the user, device, harness, session, and byte range each hit came from. From any full commit ID, trace finds the session that produced it and lets an agent read the conversation behind the change. The same capabilities ship as MCP tools and a messaging skill, so agents recover earlier decisions without you relaying context by hand. Secrets are redacted on the device before upload and again on the server, and path rules keep chosen repositories out of the index. One machine needs no server at all; add the team server with Docker Compose, Postgres, and any S3 compatible object store when a whole team's devices should share one searchable archive over TLS. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache-2.0 licensed.

Tunnel Panel manages GRE tunnels and port forwarding on Linux servers from a web panel that previews every change before it runs and verifies the result against the kernel. Create GRE, GRETAP, IP6GRE, and IP6GRETAP tunnels and see the exact commands, unit file, and rollback before anything is applied. The far end configures itself from a pairing code, as text or QR, with the sides flipped automatically so the two ends cannot disagree by a typo. Native ICMP probing with hysteresis keeps a single dropped packet from flapping a tunnel between down and back, and when ICMP is filtered the panel knocks over TCP and says so instead of crying wolf. Forward any TCP or UDP port, or a whole range, with round-robin, source-hash, or weighted balancing across destinations, and a backend that stops answering its health knock leaves the ruleset until it recovers. Watch every connection crossing the relay, who connected, which destination took them, and how much has moved, and set traffic limits per tunnel, rule, or destination. When something breaks, diagnostics reach a verdict with evidence and a suggested fix, backed by a streaming manual probe and a path-MTU search that offers to apply its findings. The panel reads your kernel's relay limits, recommends values from the host's memory and cores, and refuses changes that would lock you out of your own server. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.

Every email your app sends while you build and test it lands in OwlMail's browser inbox instead of a customer's mailbox. Point any application's SMTP settings at OwlMail and every message is ready to open as rendered HTML, plain text, headers, or raw source. Attachments and inline images sit beside the body, and seven preset preview widths show how the message lays out on phones, tablets, and desktops before anyone real sees it. A Verify panel checks subject, body, and link hosts against expectations, while CI pipelines use the REST API with delivery cursors to wait for exactly the messages their test run produced, assert on them, and clean up afterwards. Coding agents get the same inbox read-only over MCP, so an assistant can confirm a signup email arrived without touching your mail server. When a message proves itself, relay it onwards to a real SMTP server, by hand or automatically under filtering rules, or fire webhooks with HMAC signatures and retry queues into the rest of your automation. Everything stores as plain .eml files on disk, with optional SQLite indexing, S3 attachment storage, and ZIP exports for archiving. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

If your app asks an LLM the same routing, triage, or moderation question thousands of times a day, stuntd is the local proxy that learns those answers and stops paying for them. It sits in front of OpenAI or Anthropic as a drop-in replacement for the provider base URL, recording typed requests while relaying everything else byte for byte, and distils each decision site into a small head on a frozen Laya encoder. Once trained, the head answers in about 20 milliseconds with no tokens and no round trip; when its calibrated confidence falls below the site threshold, or the novelty gate flags the input, the question goes back to the provider instead of being guessed at. It speaks four protocols on one daemon: Jev System One with choice, score, and noul questions, the OpenAI Decisions API, Chat Completions, and the Anthropic Messages API, and every response carries an X-Stuntd header showing who answered. Run it keyless and the base Laya checkpoint answers zero-shot locally, put the paid Jev API in front to teach from real traffic, or import labelled rows and train directly. A few thousand rows close the gap: in the banking demo, the risk question climbs from 30.0 to 72.5 percent. A shipped Claude Code PreToolUse hook answers "may this command run" in about 50 milliseconds, without sending every command to a provider. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache-2.0 licensed.

Imp runs large AI chat models privately on one NVIDIA RTX 50 or RTX PRO graphics card, exposing them through OpenAI and Anthropic compatible APIs plus a built-in browser chat. Start it with a single Docker command, point it at a model, and the chat opens in your browser with saved conversations and a live readout of generation speed. Coding agents such as Claude Code and Codex connect by changing one base URL, so your editor talks to the model on your own hardware instead of a rented cloud endpoint. Tool calls, structured JSON output, streaming responses, and reasoning models all work over the wire, and requests can be protected with an API key, per-IP rate limits, and concurrency caps. Load GGUF files you already have or NVFP4 and FP8 SafeTensors checkpoints straight from Hugging Face with no conversion step before the first answer. When a model is bigger than the card, mixture-of-experts weights stay in system RAM while the busiest experts are cached on the GPU, so a 123.6 GiB model still answers from a 32 GB card. Supported models can also read images alongside text, and tokenize, health, and model listing endpoints make the server easy to wire into scripts and monitors. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

DocsMint is a self-hosted knowledge workspace where your team writes documentation once and both people and AI agents retrieve it through hybrid search and a built-in MCP server. Documents live in a rich visual editor with Markdown import and export, organized by folders, categories, and tags, with sharing and version history. Search blends exact title matches, multilingual keyword search, typo-tolerant fuzzy matching, and semantic vectors, then widens results through graph relationships between documents, so a query in one language or phrasing still finds the right page. Every edit quietly refreshes the index in the background; nothing needs manual re-embedding. Agents connect through 22 MCP tools covering search, reading, creation, updates, folders, snapshots, and export, over a local stdio bridge or the HTTP endpoint on your own installation. Category-scoped API keys hand each integration only the slice of the library it may read, edit, or write, while a global key serves trusted automation. The same operations are available through the REST API, a typed TypeScript SDK, and a terminal CLI for scripts. The web application installs as a PWA with offline reading, and AI features run on your choice of an OpenRouter key or a fully local Ollama setup. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache-2.0 licensed.

Hatchdoor gives your Markdown notes two front doors at once: a fast web interface for you, and an MCP endpoint where AI agents read and edit the very same vault. Point it at any folder of Markdown files, an existing Obsidian vault included, and it indexes everything into a disposable SQLite read model that rebuilds itself from your files if deleted. In the browser you browse folders, follow wikilinks that survive renames, explore backlinks and graph views, and search by keyword or meaning across every vault you mount. The live editor shows Markdown syntax only on the line you are editing and saves automatically, while concurrency checks refuse a save when the note changed underneath you. Agents connect with their own token and start read-only; turn writes on and Claude Code, Codex, Cursor, or Hermes can create, edit, move, and link notes through the same checked operations the interface uses. Saved queries in Obsidian's Bases syntax render as live tables, attachments preview inline, and optional git integration commits and pushes every change back to your vault's repository. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.

Better Status Page watches your services around the clock and turns outages into calm, branded status updates for your users instead of a flood of support tickets. Its monitors probe HTTPS endpoints with authentication flows and TLS expiry warnings, ping and TCP ports, DNS records, Docker containers, mail servers over SMTP, IMAP, and POP3, and databases from PostgreSQL and MySQL to Redis, MongoDB, SQL Server, and Oracle. A slower-than-usual alert fires when a monitor's p95 response time climbs against its own last seven days, and passive webhook heartbeats catch jobs that simply stop checking in. Alerts reach the team through email, Slack, Discord, Microsoft Teams, Telegram, ntfy, Pushover, webhooks, and Twilio SMS or voice calls, with failure thresholds, quiet hours, rate caps, and grouping keeping the noise down. Incidents carry severity levels, public updates, linked monitors, and recurring maintenance windows, and subscribers follow along by email, webhook, RSS or Atom feed, or a JSON status API. A drag-and-drop builder restyles the public page with your branding, dark mode, and translations, a static fallback keeps the page readable if the server itself falls over, and uptime reports export to CSV for SLA reviews. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

A clinic day starts before the first patient arrives, and OpenVPM has the whole of it on one screen: the schedule, the charts, the invoices, and the shelves. The calendar gives every doctor a column across day and week views, color-codes appointment types, flags doctor and room conflicts, and walks each visit from Scheduled to Checked In, In Exam, and Checked Out. Any chart opens in seconds through fuzzy search, carrying species, breed, weight trend charts, microchip numbers, photos, and allergy alerts that surface before anyone prescribes. SOAP notes sit beside problem lists, vaccination records with reminders and downloadable certificates, lab results against reference ranges, and prescriptions with a dosing calculator and refill tracking. Checkout ties the medicine to the money: treatment templates draft itemized invoices with tax, estimates convert to invoices, payments post to the account balance, and stock deducts itself when products are dispensed. Owners get their own portal for health records, appointment requests, vaccination certificates, invoices, and active prescriptions. The practice whiteboard refreshes itself as animals move between rooms, so the whole team always knows who is waiting, who is in exam, and who is done. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.

Buyers now ask an AI assistant before they ever visit your website, and Cited tells you whether your brand made it into the answer. Verify your domain with a DNS record, choose the buyer-like prompts worth watching, and Cited checks them on schedule across ChatGPT, Claude, Gemini, Perplexity, Google AI Mode, and Google AI Overviews. Every meaningful appearance becomes a durable evidence note in your citation inbox: open one to see the exact prompt, the response snapshot, where your domain appeared, and the first-seen history. It distinguishes a true source citation from a brand mention without a link, an explicit recommendation, and the uncomfortable one, a missed opportunity where a competitor got cited and you did not. Competitor tracking runs continuously, so you see rivals earning answers in your category while there is still time to respond. Notes can be annotated, organized through new, seen, saved, and resolved states, and exported for reporting. Email alerts flag meaningful new events and a weekly digest summarizes what changed, so checking becomes a habit instead of another dashboard to babysit. Workspace roles and local credentials keep a small marketing team organized on your own PostgreSQL database. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.

Started as an ExpenseOwl-inspired project, Wally grew into a lightweight, self-hosted expense tracker more capable in every direction its author touched. The backend is Python FastAPI over SQLite, which means every deployment ships a full REST API with interactive documentation at /api/docs - automating imports or wiring in external tools requires no reverse engineering, and when the optional login page is enabled you can mint scoped API keys from the Settings page for token-based integrations. The transactions view is built on AG Grid, bringing real search, column sorting, and per-column filtering to your ledger, with a footer totaling rows, income, and expenses for whatever slice you have filtered. Dashboards go beyond the usual monthly doughnut: a Change button swaps in year-scale line graphs so you can track a single category - restaurants, say - across time. Recurring transactions edit intelligently, letting you apply changes to all instances or only future ones. CSV import and export use a simple six-column format handled from Settings, the refined dark theme is genuinely easy on the eyes, and the interface is translated into more than ten languages. The whole thing runs from one small container with a single data volume.

Built atop the TauricResearch TradingAgents framework with nearly 100,000 GitHub stars, TradingAgents GUI transforms a CLI-only multi-agent LLM stock analysis pipeline into a polished web application accessible at localhost:5000. The system deploys twelve specialized AI agents — fundamental analysts, sentiment experts, technical analysts, bull and bear researchers, a trader, risk management team, and portfolio manager — who collaboratively debate market conditions through structured LangGraph workflows before producing a final BUY, SELL, or HOLD recommendation. The interface supports ten LLM providers including OpenAI, Anthropic, Google, OpenRouter, DeepSeek, Ollama, xAI, Qwen, GLM, and MiniMax, with a first-run wizard that auto-detects configured API keys and tests connections. A live pipeline visualization shows each agent's status with real-time progress bars, while the tabbed output area separates Live Feed, Reports preview, and Tool calls into dedicated panes. The three-pane Reports tab provides searchable indexing, table-of-contents navigation, and export to Markdown, HTML, or PDF formats. Report length control across Concise, Standard, and Comprehensive modes saves up to 50% on token costs. Multi-session chat allows pinning past reports as grounding context with live token counting and context-window warnings. Three built-in themes — Terminal, Modern, and Bloomberg — persist per browser. Docker Compose deployment maps port 5000 with persistent report storage. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache 2.0 licensed.

Email discussion lists - the venerable listserv pattern where one address relays to a whole group - without the notorious setup pain of Mailman or Sympa: that is Gray Duck Mail. Its architectural trick is requiring no mail server of your own: the C#/ASP.NET Core app monitors a POP3 inbox at any external email host supporting POP3/SMTP and aliasing, relays each incoming message to all list members via SMTP, and archives it for browsing in the web interface. Replies route back to the entire group automatically. The tedious parts of list management are handled by convention: dedicated aliases process join requests, subscription confirmations, and unsubscribes, while a bounce alias in the return-path catches delivery failures so dead addresses get handled instead of silently rotting. The web administration interface creates and removes lists, manages contacts, browses message archives, and imports or exports the local database for backup. GPLv3-licensed, Docker-distributed, and fully code-documented, it targets exactly the groups email still serves best - neighborhoods, schools, congregations, workplaces, and families - where one message keeping everyone in the loop beats yet another app nobody wants to install.

BlogFactory is the first open-source agent control plane purpose-built for multi-site content operations, shipping 16 distinct capabilities across four operational pillars — Operate, Grow, Control, and Deliver — in its v0.1.0 release. The platform keeps source evidence, AI-generated drafts, revisions, SEO metadata, human review, preflight checks, and CMS delivery destinations in one workspace while enforcing a server-side authority ceiling that prevents agents from ever publishing live content, accessing credentials, or performing administrative actions. AI agents connect through a Streamable HTTP MCP endpoint with OAuth authentication and site-scoped tool discovery, compatible with Codex, ChatGPT, and any MCP-capable client. Content creation supports single articles from keywords, titles, URLs, PDFs, text, or YouTube sources, batch campaigns, and programmatic content via reusable templates with structured data rows. The Search Growth module connects Google Search Console complete-day data to ranked optimization opportunities, 30-day growth plans, content classification by signal type (growing, low-CTR, almost-ranking, page-two, zero-click), indexing checks, and semantic internal-link suggestions. Brand Voice separates global voice rules, writer profiles, knowledge context, and brand guidelines from article mechanics. Approved drafts are delivered as CMS drafts to WordPress, Ghost, Wix, or Framer through idempotent, optimistic-locking-protected handoffs with explicit destination selection and preflight validation. The stack runs on React/Vite, a Hono/TypeScript API, PostgreSQL, and MinIO via Docker Compose, with providers including OpenRouter, OpenAI, and Google services under a bring-your-own-AI model. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.